Giovanni Trappolini
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Giovanni Trappolini has a total of 32 co-authors. Here are the top co-authors by number of shared publications:
- Fabrizio Silvestri
Sapienza University of Rome, Rome, Italy
8 shared publications - Federico Siciliano
Sapienza University of Rome, Rome, Italy
5 shared publications - Nicola Tonellotto
University of Pisa, Pisa, Italy
3 shared publications - Antonio Purificato
Sapienza University of Rome, Rome, Italy
2 shared publications - Emanuele Rodolà
Sapienza University of Rome Italy
2 shared publications - Fabio Petroni
Samaya AI, London, United Kingdom
2 shared publications
Latest Publications
- Open Access
- 1 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Control Systems Technology
- Vol. 34 (3) , 1133-1148
Adaptive Estimation of Pennes’ Bio-Heat Equation: Observer Design and PINNs-Based Implementation
This article presents a multiple-model (MM) adaptive framework for state estimation of uncertain parabolic reaction-diffusion partial differential equations, with specific application to superficial oncological hyperthermia (HT). The effectiveness of such thermal processes critically depends on real-time temperature assessment—a significant difficulty when relying on intrinsically sparse thermometry. This challenge is further compounded by uncertainties in patient-specific properties, particularly blood perfusion rate, during tissue heating. The problem is formulated as a state estimation of the governing Pennes’ bio-heat equation (PBHE), developing an MM adaptive estimation framework capable of dealing with uncertain parameters in the process dynamics and using only noninvasive boundary measurements. The validation process revealed limitations in traditional numerical simulation methods, prompting a novel implementation leveraging physics-informed learning—an approach well-suited for scenarios with known governing equations but limited training data. The experimental validation using a muscle-equivalent phantom with controlled perfusion simulation provides evidence of the method’s effectiveness in realistic conditions. This is the first demonstration of adaptive estimation of parabolic partial differential equations (PDEs) based on physiscs informed learning methods. The proposed methodology offers three key advantages: elimination of domain discretization requirements, thereby enabling continuous spatiotemporal monitoring; real-time estimation capabilities for PDE solutions; scalability to higher dimensional input spaces for an increased adaptability.
- 1 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Control Systems Technology
- Vol. 34 (3) , 1133-1148
Open AccessAdaptive Estimation of Pennes’ Bio-Heat Equation: Observer Design and PINNs-Based Implementation
This article presents a multiple-model (MM) adaptive framework for state estimation of uncertain parabolic reaction-diffusion partial differential equations, with specific application to superficial oncological hyperthermia (HT). The effectiveness of such thermal processes critically depends on real-time temperature assessment—a significant difficulty when relying on intrinsically sparse thermometry. This challenge is further compounded by uncertainties in patient-specific properties, particularly blood perfusion rate, during tissue heating. The problem is formulated as a state estimation of the governing Pennes’ bio-heat equation (PBHE), developing an MM adaptive estimation framework capable of dealing with uncertain parameters in the process dynamics and using only noninvasive boundary measurements. The validation process revealed limitations in traditional numerical simulation methods, prompting a novel implementation leveraging physics-informed learning—an approach well-suited for scenarios with known governing equations but limited training data. The experimental validation using a muscle-equivalent phantom with controlled perfusion simulation provides evidence of the method’s effectiveness in realistic conditions. This is the first demonstration of adaptive estimation of parabolic partial differential equations (PDEs) based on physiscs informed learning methods. The proposed methodology offers three key advantages: elimination of domain discretization requirements, thereby enabling continuous spatiotemporal monitoring; real-time estimation capabilities for PDE solutions; scalability to higher dimensional input spaces for an increased adaptability.
- Open Access
- 18 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 147-154
Eclipse: Contrastive Dimension Importance Estimation with Pseudo-Irrelevance Feedback for Dense Retrieval
- 18 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 147-154
Open AccessEclipse: Contrastive Dimension Importance Estimation with Pseudo-Irrelevance Feedback for Dense Retrieval
- Open Access
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 4168-4171
IR-RAG @SIGIR25: The Second Edition of the Workshop on Information Retrieval's Role in RAG Systems
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 4168-4171
Open AccessIR-RAG @SIGIR25: The Second Edition of the Workshop on Information Retrieval's Role in RAG Systems
- Open Access
- 8 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 82668-82679
Quantized Auto Encoder-Based Anomaly Detection for Multivariate Time Series Data in 5G Networks
With the arrival of 5G technology, networks face critical challenges in detecting anomalies that can significantly impact performance and reliability. This paper introduces QAED (Quantized Auto Encoder Detector), a novel deep learning approach for anomaly detection in 5G networks with three key innovations: (1) a vector quantization mechanism that effectively captures discrete network states, (2) a kernel density estimation preprocessing step that enables detection of both outliers and distribution shifts, and (3) an integrated architecture that processes multivariate time series data in a unified framework. We provide a detailed evaluation of our model across 5G data scenarios, demonstrating its enhanced accuracy and efficiency in anomaly detection compared to existing state-of-the-art methods, with gains of up to 8%.
- 8 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 82668-82679
Open AccessQuantized Auto Encoder-Based Anomaly Detection for Multivariate Time Series Data in 5G Networks
With the arrival of 5G technology, networks face critical challenges in detecting anomalies that can significantly impact performance and reliability. This paper introduces QAED (Quantized Auto Encoder Detector), a novel deep learning approach for anomaly detection in 5G networks with three key innovations: (1) a vector quantization mechanism that effectively captures discrete network states, (2) a kernel density estimation preprocessing step that enables detection of both outliers and distribution shifts, and (3) an integrated architecture that processes multivariate time series data in a unified framework. We provide a detailed evaluation of our model across 5G data scenarios, demonstrating its enhanced accuracy and efficiency in anomaly detection compared to existing state-of-the-art methods, with gains of up to 8%.
- Open Access
- 24 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 5762-5772
A Reproducible Analysis of Sequential Recommender Systems
Sequential Recommender Systems (SRSs) have emerged as a highly efficient approach to recommendation systems. By leveraging sequential data, SRSs can identify temporal patterns in user behaviour, significantly improving recommendation accuracy and relevance. Ensuring the reproducibility of these models is paramount for advancing research and facilitating comparisons between them. Existing works exhibit shortcomings in reproducibility and replicability of results, leading to inconsistent statements across papers. Our work fills these gaps by standardising data pre-processing and model implementations, providing a comprehensive code resource, including a framework for developing SRSs and establishing a foundation for consistent and reproducible experimentation. We conduct extensive experiments on several benchmark datasets, comparing various SRSs implemented in our resource. We challenge prevailing performance benchmarks, offering new insights into the SR domain. For instance, SASRec does not consistently outperform GRU4Rec. On the contrary, when the number of model parameters becomes substantial, SASRec starts to clearly dominate all the other SRSs. This discrepancy underscores the significant impact that experimental configuration has on the outcomes and the importance of setting it up to ensure precise and comprehensive results. Failure to do so can lead to significantly flawed conclusions, highlighting the need for rigorous experimental design and analysis in SRS research. Our code is available at https://github.com/antoniopurificato/recsys_repro_conf.
- 24 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 5762-5772
Open AccessA Reproducible Analysis of Sequential Recommender Systems
Sequential Recommender Systems (SRSs) have emerged as a highly efficient approach to recommendation systems. By leveraging sequential data, SRSs can identify temporal patterns in user behaviour, significantly improving recommendation accuracy and relevance. Ensuring the reproducibility of these models is paramount for advancing research and facilitating comparisons between them. Existing works exhibit shortcomings in reproducibility and replicability of results, leading to inconsistent statements across papers. Our work fills these gaps by standardising data pre-processing and model implementations, providing a comprehensive code resource, including a framework for developing SRSs and establishing a foundation for consistent and reproducible experimentation. We conduct extensive experiments on several benchmark datasets, comparing various SRSs implemented in our resource. We challenge prevailing performance benchmarks, offering new insights into the SR domain. For instance, SASRec does not consistently outperform GRU4Rec. On the contrary, when the number of model parameters becomes substantial, SASRec starts to clearly dominate all the other SRSs. This discrepancy underscores the significant impact that experimental configuration has on the outcomes and the importance of setting it up to ensure precise and comprehensive results. Failure to do so can lead to significantly flawed conclusions, highlighting the need for rigorous experimental design and analysis in SRS research. Our code is available at https://github.com/antoniopurificato/recsys_repro_conf.
- Open Access
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 3036-3039
IR-RAG @ SIGIR24: Information Retrieval's Role in RAG Systems
In recent years, Retrieval Augmented Generation (RAG) systems have emerged as a pivotal component in the field of artificial intelligence, gaining significant attention and importance across various domains. These systems, which combine the strengths of information retrieval and generative models, have shown promise in enhancing the capabilities and performance of machine learning applications. However, despite their growing prominence, RAG systems are not without their limitations and continue to be in need of exploration and improvement. This workshop seeks to focus on the critical aspect of information retrieval and its integral role within RAG frameworks. We argue that current efforts have undervalued the role of Information Retrieval (IR) in the RAG and have concentrated their attention on the generative part. As the cornerstone of these systems, IR's effectiveness dramatically influences the overall performance and outcomes of RAG models. We call for papers that will seek to revisit and emphasize the fundamental principles underpinning RAG systems. At the end of the workshop, we aim to have a clearer understanding of how robust information retrieval mechanisms can significantly enhance the capabilities of RAG systems. The workshop will serve as a platform for experts, researchers, and practitioners. We intend to foster discussions, share insights, and encourage research that underscores the vital role of Information Retrieval in the future of generative systems.
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 3036-3039
Open AccessIR-RAG @ SIGIR24: Information Retrieval's Role in RAG Systems
In recent years, Retrieval Augmented Generation (RAG) systems have emerged as a pivotal component in the field of artificial intelligence, gaining significant attention and importance across various domains. These systems, which combine the strengths of information retrieval and generative models, have shown promise in enhancing the capabilities and performance of machine learning applications. However, despite their growing prominence, RAG systems are not without their limitations and continue to be in need of exploration and improvement. This workshop seeks to focus on the critical aspect of information retrieval and its integral role within RAG frameworks. We argue that current efforts have undervalued the role of Information Retrieval (IR) in the RAG and have concentrated their attention on the generative part. As the cornerstone of these systems, IR's effectiveness dramatically influences the overall performance and outcomes of RAG models. We call for papers that will seek to revisit and emphasize the fundamental principles underpinning RAG systems. At the end of the workshop, we aim to have a clearer understanding of how robust information retrieval mechanisms can significantly enhance the capabilities of RAG systems. The workshop will serve as a platform for experts, researchers, and practitioners. We intend to foster discussions, share insights, and encourage research that underscores the vital role of Information Retrieval in the future of generative systems.
- Top Cited PapersOpen Access
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 719-729
The Power of Noise: Redefining Retrieval for RAG Systems
Retrieval-Augmented Generation (RAG) has recently emerged as a method to extend beyond the pre-trained knowledge of Large Language Models by augmenting the original prompt with relevant passages or documents retrieved by an Information Retrieval (IR) system. RAG has become increasingly important for Generative AI solutions, especially in enterprise settings or in any domain in which knowledge is constantly refreshed and cannot be memorized in the LLM. We argue here that the retrieval component of RAG systems, be it dense or sparse, deserves increased attention from the research community, and accordingly, we conduct the first comprehensive and systematic examination of the retrieval strategy of RAG systems. We focus, in particular, on the type of passages IR systems within a RAG solution should retrieve. Our analysis considers multiple factors, such as the relevance of the passages included in the prompt context, their position, and their number. One counter-intuitive finding of this work is that the retriever's highest-scoring documents that are not directly relevant to the query (e.g., do not contain the answer) negatively impact the effectiveness of the LLM. Even more surprising, we discovered that adding random documents in the prompt improves the LLM accuracy by up to 35%. These results highlight the need to investigate the appropriate strategies when integrating retrieval with LLMs, thereby laying the groundwork for future research in this area.
- 10 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 719-729
Top Cited PapersOpen AccessThe Power of Noise: Redefining Retrieval for RAG Systems
Retrieval-Augmented Generation (RAG) has recently emerged as a method to extend beyond the pre-trained knowledge of Large Language Models by augmenting the original prompt with relevant passages or documents retrieved by an Information Retrieval (IR) system. RAG has become increasingly important for Generative AI solutions, especially in enterprise settings or in any domain in which knowledge is constantly refreshed and cannot be memorized in the LLM. We argue here that the retrieval component of RAG systems, be it dense or sparse, deserves increased attention from the research community, and accordingly, we conduct the first comprehensive and systematic examination of the retrieval strategy of RAG systems. We focus, in particular, on the type of passages IR systems within a RAG solution should retrieve. Our analysis considers multiple factors, such as the relevance of the passages included in the prompt context, their position, and their number. One counter-intuitive finding of this work is that the retriever's highest-scoring documents that are not directly relevant to the query (e.g., do not contain the answer) negatively impact the effectiveness of the LLM. Even more surprising, we discovered that adding random documents in the prompt improves the LLM accuracy by up to 35%. These results highlight the need to investigate the appropriate strategies when integrating retrieval with LLMs, thereby laying the groundwork for future research in this area.
- 5 July 2024
- Published by Springer Nature
- p. 285-291
TEXT2TASTE: A Versatile Egocentric Vision System for Intelligent Reading Assistance Using Large Language Model
- 5 July 2024
- Published by Springer Nature
- p. 285-291
TEXT2TASTE: A Versatile Egocentric Vision System for Intelligent Reading Assistance Using Large Language Model
- 26 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. 5 (5) , 2293-2303
Sparse Vicious Attacks on Graph Neural Networks
In this study, we introduce SAVAGE, a novel framework for sparse vicious adversarial link prediction attacks in Graph Neural Networks (GNNs). While GNNs have been successful in link prediction tasks, they are susceptible to adversarial attacks where malicious nodes attempt to manipulate recommendations for a target victim. SAVAGE optimizes the attacker's goal to maximize attack effectiveness while minimizing the required malicious resources. Unlike existing methods with static resource-based upper bounds, SAVAGE employs a sparsity enforcing mechanism to reduce the number of malicious nodes needed for the attack. Extensive experiments on real-world and synthetic datasets demonstrate the optimal trade-off achieved by SAVAGE between a high attack success rate and the number of malicious nodes utilized. Furthermore, we demonstrate that SAVAGE can successfully target non-GNN-based link prediction systems, even those unknown at the time of the attack. This showcases the transferability of SAVAGE-generated attacks to other black-box methods for link prediction, highlighting its applicability across different real-world scenarios.
- 26 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. 5 (5) , 2293-2303
Sparse Vicious Attacks on Graph Neural Networks
In this study, we introduce SAVAGE, a novel framework for sparse vicious adversarial link prediction attacks in Graph Neural Networks (GNNs). While GNNs have been successful in link prediction tasks, they are susceptible to adversarial attacks where malicious nodes attempt to manipulate recommendations for a target victim. SAVAGE optimizes the attacker's goal to maximize attack effectiveness while minimizing the required malicious resources. Unlike existing methods with static resource-based upper bounds, SAVAGE employs a sparsity enforcing mechanism to reduce the number of malicious nodes needed for the attack. Extensive experiments on real-world and synthetic datasets demonstrate the optimal trade-off achieved by SAVAGE between a high attack success rate and the number of malicious nodes utilized. Furthermore, we demonstrate that SAVAGE can successfully target non-GNN-based link prediction systems, even those unknown at the time of the attack. This showcases the transferability of SAVAGE-generated attacks to other black-box methods for link prediction, highlighting its applicability across different real-world scenarios.
- Open Access
- 18 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 2619-2628
Multimodal Neural Databases
- 18 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 2619-2628
Open AccessMultimodal Neural Databases